What you don’t want is more ivory tower geeks writing in lisp who obsesses about algorithmic purity and can produce noddy systems that work for some cases but will break badly in the real world.
One example from the 80's I was helping build billing systems for BT I also had to run the system and make sure it all worked properly. The first time we hit £1,000,000 in a month (about $5,000,000) in today’s money ) we had a small celebration and I recall the CTO (one of Vints reports I believe) nudging me and saying "this had better be right or we are both out of a Job"
Its like saying we need more engineers like Ross Brawn I know lets train more physicists 10 years later you have 200 Brian Cox's and not 195 Ross Brawns and 5 Brian Cox's working at Cern